ASTER-BASED LITHOLOGIC AND ALTERATION MAPPING IN KHETRI COPPER BELT, INDIA

Abstract

The thesis presents a detailed account of the analysis of an ASTER dataset of a mineralized region in northwest India, the Khetri Copper Belt. ASTER acquires data in the geologically important wavelength regions of the electromagnetic spectrum, namely the visible-nearinfrared- through-shortwave-infrared (VNIR-SWIR; 0.5-2.5 um), and the thermal-infrared (TIR; 8.0-11.5 Jim) in 14 discrete strategically selected spectral bands. Most alteration minerals related to various kinds of mineralization have characteristic spectral signatures in the solar reflective VNIR-SWIR region. Important rock forming silicate minerals, like quartz and feldspar, have diagnostic spectral features in the TIR wavelength region. ASTER data of the study area in these two spectral regions have been digitally processed to map the lithologies and surface/alteration mineral assemblages, using existing and improved methods of image processing. The TIR data has been processed to map lithology and the VNIR-SWIR data has been used to map surface/alteration minerals. KCB is a structurally complexand regionally metamorphosed NNE-SSW striking linear polymetallic-sulfide mineralized belt consisting primarily of early Proterozoic metasedimentary rock units, and late Proterozoic basic and acidic intrusives. The older psammitic Alwar Group (mainly pure and impure quartzites, and micaceous quartzites), and younger pelitic Ajabgarh Group (mainly phyllites, schists, impure marbles and calc-silicate units) comprise the main lithostratigraphic units of KCB; along with many small stocks, sills and dikes of granite, dolerite and amphibolite. The study area is marked by a prominent NNE-SSW striking cataclastic/shear zone along which numerous felsic (granitic and pegmatitic) intrusives are emplaced. Small, but significant sulfide mineralization, and associated rock alteration close to the surface in a few places and in vicinity of these intrusives are observed. Previous TM-based remote sensing studies in a part of the study area have indicated anomalous zones of OH" alteration. This study has been designed to investigate the extent to which ASTER can improve II the existing understanding of the alteration patterns of the area and the associated lithologic units. Ground truth in the form of fieldwork, and laboratory analyses (thin-section petrography and trace-metal geochemistry) of selected field samples have been used to validate the results ofASTER data analysis. Retrieval of reflectance and emissivity information from the at-sensor radiance data has involved an assessment of various radiometric, atmospheric and topographic correction methods and selecting the best technique for use in spectral processing. Best atmospheric correction for the 9-band VNIR-SWIR data has been achieved using a hybrid approach which is based on enhancement of MODTRAN-based atmospheric correction through Modified Flat- Field correction of the result. This has served as the basic data in spectral processing for surface/alteration mapping objective. The standard Level 2 surface emissivity product has been used in the detailed spectral processing for the lithologic mapping objective. ASTER TIR at-sensor radiance and surface emissivity data have been used to produce the surface lithologic maps of the study area. Decorrelation stretch processing and lithologic indices calculated using the TIR radiance data for quartz-rich and mafic-rich rocks have provided best qualitative lithologic discrimination. Full spectral processing of the surface emissivity data has produced a lithologic map of the area with the major lithologic units classified into six classes: namely, 'mica schist', 'felsic granite', 'mafic diabase', 'pink quartzite', 'brown to dark brown sand', and 'brown to dark brown sandy loam', based on the matches of the image-derived end-member spectra with the reference spectral library spectra (Johns Hopkins University spectral library). Two approaches for quantitative silica abundance estimation have been investigated. Method 1 is based on the spectral modeling of the surfaceemissivity spectra; whereas Method 2 is based on Metal Mining Agency of Japan's (MMAJ) original 'K-value' method. The two results have subsequently been compared in specific context of the study area. Results of the analysis indicate a general overall correspondence with the reported and mapped lithologic units of the study area. Ill The VNIR-SWIR reflectance data have been used to generate surface/alteration mineral distribution maps for the study area. A number of minerals have been mapped through different approaches, which involved use of band ratios, relative absorption band depth images (RBDs), Feature-oriented Principal Components Selection (FPCS) technique, spectral indices, and Boolean logical operator-based alteration group mapping. The full spectral processing of the reflectance data has helped in unique identification and mapping of a variety of surface/alteration mineral species based on spectral matches with a reference spectral library spectra (United States Geological Survey spectral library; speclib05). Validation of the results of ASTER-based spectral lithologic and alteration mapping based on the ground-truth has revealed that despite its multispectral nature, ASTER data can provide valuable and reliable surface lithologic and mineral maps which can be directly used in prospecting for mineral deposits.